Azure OpenAI Integration for AI Governance Projects

You can integrate an instance of Azure OpenAI with OneTrust to create project records in OneTrust based on deployments in Azure OpenAI Service. This offers Data Scientists a workstream that is more efficient than manual data entry.

Prerequisites

  1. Azure subscription + access to Azure OpenAI

  2. Active deployments in Azure AI Studio

  3. Azure credential in OneTrust

  4. Azure Endpoints

  5. Azure Keys (API keys)

To obtain endpoints and API keys in Azure

  1. In the Azure AI Studio console, navigate to your resource group.
  2. In the Keys and Endpoint menu, copy the keys and endpoint details to use in the credential configuration.

    Microsoft Azure - Keys and Endpoint

To create Azure credential

  1. In the Integrations menu, select Credentials.

    The Credentials screen appears.

  2. Click the Add New button.

    The Select System screen appears.

  3. Choose Azure OpenAI Service from the Select System field.

    Select system screen
  4. Click the Next button.

    The Enter Credential Details screen appears.

  5. Provide the credential a unique name, then choose Web App as the Connector type and API Key as the Auth Type.

    Enter credential details
  6. Provide the Endpoint from Azure and in Hostname field.
  7. Complete in the Key and Value fields with the API key and Value copied from Azure Keys.
  8. Set Authentication Parameters to Send in Headers.
  9. Click the Save button.

To create the integration

  1. In the Integrations menu, select Gallery.

    The Integration Gallery screen appears.

  2. Use the search function to locate the Azure OpenAI tile.

    Integrations gallery Azure OpenAI
  3. Click the tile.

    The Gallery Details screen appears.

  4. Click the Add button for Azure OpenAI Service Deployment Registration.

    The Add Workflow Name modal appears.


    Azure OpenAI Service Deployment Registration tile
  5. Complete the required fields. Ensure to click the Create button.

    The Integration Builder screen appears.


    Integration builder - Add workflow name
  6. The integration will have a scheduler trigger that you can further configure for how often you would like to run the workflow.

    Workflow builder screen - Edit integration
  7. The integration will include a seeded action to look for deployments using your chosen Azure credential. The request URL is created using the endpoint from your Azure instance that is provided in the credential setup.

    Azure OpenAI Request URL
  8. The POST create action will leverage the Apply to Each logic to create inventory records for each deployment.
    Note:

    Integration workflows can only have one Apply to Each logic action per workflow. This is designed to avoid excessive processing time, as workflows processing for more than 25 minutes will fail. OneTrust recommends to send a second Apply To Each loop to process in a subflow workflow.


    Azure OpenAI post creation
  9. Save and activate the integration. Whenever a deployment is created and registered in Azure AI, a new project inventory record will be created. This inventory will also feature the relationship record in the Related tab.